In normal hearing, sound signals are converted into electrical impulses that are perceived by the brain as sound. Conditions which affect the auditory system can result in a range of hearing impairments. The perception of auditory signals that are received by a hearing impaired patient may be improved by prosthetic implants that incorporate a microphone, a signal processor and an electrode array for delivering representative electrical signals to the patient. The signal processor converts auditory signals into electrical signals in order to convey the sound to the patient. Of particular importance to patients having a hearing deficiency is in regard to the comprehension of speech.
Some conditions can be overcome by hearing aids and result in near normal hearing and good speech understanding. However, some conditions which lead to more severe hearing loss cannot be overcome with conventional hearing aids. These pathologies can be overcome by implanting auditory neural prostheses which bypass the damaged part of the auditory system and typically activate the auditory system through electrical stimulation to provide the perception of sound. Cochlear implants, auditory brainstem implants and auditory midbrain implants are all examples of such auditory neural prostheses.
Auditory signals (sounds) may be described as having various elements such as loudness (amplitude) and pitch as well as varying over time. Each of these parameters must be considered when developing an auditory signal processing method. A signal processing method is a predetermined instruction set for producing electrode stimulus instructions from received sound signals.
The nature of the normal function of hearing has previously been described (U.S. Pat. No. 5,271,397). In particular, when normal speech is analysed it is found that several frequency peaks are produced simultaneously. This provides the nature of the sound and the characteristics of speech interpretation. The frequency peaks, known as formants, are numbered from the glottal pulse, F0, with higher frequency peaks being the first formant, F1, second formant, F2 and so on. Different vowel sounds change the frequency and amplitude of these formants, and in particular, the second formant F2.
In an auditory prosthesis, there are two basic methods by which neural information may be coded, the rate code and the temporal code. The rate code uses the number of neural firing events over a short time period to code auditory features. The temporal code uses the temporal position of firing events of each neuron to code auditory features. In an auditory prosthesis, the rate code is transmitted using the power of the incoming stimuli; low power electrical stimuli produce lower electrical stimulation rates, higher power electrical stimuli produce higher rates. On the other hand, the temporal code could be achieved in an auditory prosthesis by the precise presentation of electrical stimuli in time.
In regards to the pitch aspect of the signal processing strategy, the frequency is measured in Hertz (Hz), which may vary from low sounds to high sounds in a range from 20 Hz to 20,000 Hz. The auditory signal processing method divides the auditory signal into bands of frequencies, one method being to divide the signal into frequency bands approximating a quarter of an octave (PCT/AU00/00838).
Alternatively, a spatio-temporal pattern of stimulation along the length of an intra-cochlea electrode array may be produced which delays more apical stimuli (PCT/AU01/00723). This mimics the spatio-temporal pattern associated with the travelling wave observed on the basilar membrane in an acoustically excited normally-hearing cochlea. Although this strategy mimics a known physiological process adding a fixed delay to each electrode to compensate for bypassed processes, it essentially carries no additional information regarding the incoming signal. It merely changes the time that the information is provided through the electrode.
Simultaneous stimulation of the electrodes is not conducive to eliciting a perception of sound that is faithful to the actual incoming acoustic signal. This is because if electrodes are stimulated simultaneously, current paths between electrodes can interact, causing undesirable stimulation. Therefore, most existing cochlear strategies have been developed to stimulate only one electrode at a time.
The aspect of a signal processing strategy that relates to the variation of an auditory signal over time is known as temporal variation. It has previously been known to divide an incoming auditory signal into discrete time periods known as “timeslices” in order that the corresponding electrical signal from each timeslice may be delivered by a stimulation pulse to an appropriate electrode within the auditory prosthesis. Each timeslice incorporates the total time taken to receive, process, deliver and recover from the stimulus.
Various auditory signal processing methods have previously been described, such as CIS (U.S. Pat. No. 4,207,441) and SPEAK (U.S. Pat. No. 5,597,380). In these methods, the duration of each timeslice is fixed to a predetermined rate. In another strategy, described in PCT/AU00/00838, the stimulation rate is not fixed, but determined according to attributes contained within the incoming auditory signal.
Alternatively, the time of stimulus of each electrode may be synchronised to the temporal peak in the filter output of the sound signal. The time of stimulation being set to stimulate at the time that positive peaks occur in each frequency band (AU 2002312636).
A temporal adjustment may also be made to the electrode activation time such that the activation of lower amplitude components of the signal is delayed relative to activation of higher amplitude components of the incoming sound signal (PCT/AU2004/001729).
These auditory signal processing methods such as PCT/AU03/00639 (STAR) include generating a series of electrical stimulation “spikes” from each sound signal, where each spike has a temporal position based on the time at which the sound signal crossed a pre-determined threshold in a positive direction. The value of the threshold is adjustable in order to take account of differing listening conditions and levels of background noise. Explicit extraction of pitch is not required to control the rate of stimulation. This strategy adds additional information to the signal through the varied rate of stimulation that is used on each electrode, since stimulus rate does affect frequency precepts. Although some temporal information is added to the signal through this process, it is expected to be limited since stimuli are independently derived from the band pass signal in each electrode channel and not relative to each other.
Other auditory signal processing methods have been developed with a view to modifying the electrical stimulation of the electrodes to reflect the natural delay which occurs between an auditory signal reaching one ear and the other (interaural time delay). For instance, PCT/AU02/00660 describes the synchronization of the timing of the electrode stimuli with the temporal peak in the amplitude of the corresponding incoming band-pass filtered auditory signal. This provides advantages to patients having an auditory prosthesis in each ear, conveying the time-of-arrival differences between the ears.
Despite the progress made by existing auditory prostheses, signal processors and signal processing strategies, the perception of sound by hearing impaired patients remains imperfect and problematic. None of the temporal strategies described above have a global time point to which electrical stimuli are referenced. Furthermore, these strategies may involve generation of continuous stimulation, not allowing a particular sequence of pulses to be processed by the brain before the next sequence of pulses is delivered. It is therefore desirable to provide improved signal processing methods, auditory prostheses and systems.